Efficient Catalyst for Acceptorless Alcohol Dehydrogenation: Interplay

Feb 10, 2014 - Fukui Institute for Fundamental Chemistry, Kyoto University, Takano-Nishihiraki-cho 34-4, Sakyo-ku, Kyoto 606-8103, Japan. ‡Graduate ...
0 downloads 9 Views 4MB Size
Download PDF
Research Article pubs.acs.org/acscatalysis

Efficient Catalyst for Acceptorless Alcohol Dehydrogenation: Interplay of Theoretical and Experimental Studies Guixiang Zeng,† Shigeyoshi Sakaki,*,† Ken-ichi Fujita,*,‡ Hayato Sano,‡ and Ryohei Yamaguchi*,‡ †

Fukui Institute for Fundamental Chemistry, Kyoto University, Takano-Nishihiraki-cho 34-4, Sakyo-ku, Kyoto 606-8103, Japan Graduate School of Human and Environmental Studies, Kyoto University, Yoshida-Nihonmatsu-cho, Sakyo-ku, Kyoto 606-8501, Japan

‡

S Supporting Information *

ABSTRACT: The promoterless AAD (acceptorless alcohol dehydrogenation) reaction mediated by an iridium catalyst Cp*Ir(bpyO) 1−Ir (Cp* = pentamethylcyclopentadienyl, bpyO = α,α′-bipyridonate) has been theoretically investigated with the density functional theory. The reaction occurs through three steps, including alcohol dehydrogenation, formation of dihydrogen complex, and H2 elimination from the iridium center. In the first two steps, the metal center and the bpyO ligand work cooperatively via the aromatization/dearomatization process of the bpyO ligand. The second step is ratedetermining, where the ΔG0≠ and ΔG0 values are 23.9 and 13.9 kcal/mol, respectively. Our calculations demonstrate that the aromatization of the bpyO ligand as well as the charge transfer (CT) from the Cp* ligand to the iridium center plays important roles in stabilizing the transition state of the rate-determining step. We have theoretically and experimentally examined the 4d rhodium analogue Cp*Rh(bpyO) 1−Rh and found that it exhibits similar activity to that of 1−Ir. On the basis of those results, a new catalyst (HMB)Ru(bpyO) 1−Ru (HMB = hexamethylbenzene) is designed both theoretically and experimentally, where a cheaper and more abundant 4d ruthenium element is employed with the HMB and bpyO ligands. Theoretical calculations certainly show that 1−Ru is active for the promoterless AAD reaction via the same reaction mechanism as that of the reaction by 1−Ir. The experiments also demonstrate that 1−Ru is as efficient as 1−Ir for the AAD reaction. KEYWORDS: acceptorless alcohol dehydrogenation, theoretical calculation, iridium, cooperative catalysis, rhodium, ruthenium

■

INTRODUCTION The acceptorless alcohol dehydrogenation (AAD) with the release of H2 provides promising synthetic routes for such carbonyl compounds as aldehydes, ketones, esters, and amides.1 It also exhibits potential for the H2 production from biomass and/or its fermentation products, which is crucial for the hydrogen production technology.2 Various AAD reactions have been reported since its discovery about 40 years ago.3 In most of them, acid or base promoters are required. For example, a series of the acid-promoted AAD reactions mediated by Ru(OCOCF3)2(CO)(PPh3)2 were reported by Robinson and his co-workers in the 1970s.3a,b Subsequently, a variety of basepromoted AAD reactions were reported by Morton and ColeHamilton.3c,d In these works, the reactions occur on the metal center, while the ligand acts as a spectator.4 Owing to the increasing demand for environmentally benign synthetic processes, promoterless AAD reactions are desirable. In recent years, the cooperative catalysts involving combined basic and acidic sites were found to satisfy this requirement. However, examples of the cooperative catalytic cycle have been limited so far. One good example is a series of ruthenium pincer complexes (Cat1 to Cat3) developed by Milstein’s group,5 which catalyze the AAD reactions under neutral reaction conditions; see Scheme 1. A cationic cobalt(II) alkyl complex reported by Hanson’s group was also active for the AAD reaction without additives.6 However, high temperatures (110− © 2014 American Chemical Society

Scheme 1. Ruthenium-Based Catalysts for the AAD Reaction Developed by Milstein′s Group

120 °C) are needed in these reactions. Since 2007, several iridium complexes (Cat4 to Cat6) have been reported by Yamaguchi, Fujita, and their co-workers, which catalyze the AAD reactions efficiently under mild conditions (reflux in toluene or water), as shown in Scheme 2.7a−e In addition, similar catalysts have been reported for transfer hydrogenation, water oxidation, and carbon dioxide hydrogenation.7f−j Besides the experimental work, a number of computational works have Received: November 24, 2013 Revised: January 30, 2014 Published: February 10, 2014 1010

dx.doi.org/10.1021/cs401101m | ACS Catal. 2014, 4, 1010−1020

ACS Catalysis

Research Article

the activity of the catalyst decreases substantially by the substitution of phenO for bpyO. On the basis of these results, we proposed new AAD catalysts by theoretical work and evaluated their catalytic activities by experiments.

Scheme 2. Iridium Complexes Employed for the AAD Reaction

■

RESULTS AND DISCUSSION AAD Reaction of Benzyl Alcohol by Cp*Ir(bpyO) 1− Ir.9 There are two possible reaction pathways for the AAD reaction; one is the outer-sphere pathway and another is the inner-sphere β-H elimination one.10 Here, we employed the benzyl alcohol 2 as a substrate in calculations. Outer-Sphere Pathway. In the outer-sphere concerted pathway (see Scheme 3), the AAD reaction occurs by the cooperation of the iridium center and the bpyO ligand, which directly affords a hydride complex Cp*Ir(H)(pyOpyOH) 4−Ir (pyOpyOH = α-pyridonate-α′-hydroxylpyridine) through a transition state TS1/4−Ir, as shown in Figure 1. In TS1/4−Ir, the H1 atom migrates from the hydroxyl group (O3−H1) of 2 to the carbonyl O2 atom of the bpyO ligand, where the O3−H1 distance increases to 1.295 Å and the O2−H1 distance decreases to 1.120 Å. At the same time, the H2 atom migrates to the iridium center from 2, where the Ir−H2 distance shortens to 1.682 Å and the C2−H2 distance lengthens to 1.524 Å. These geometrical features suggest that the O3−H1 and C2−H2 bonds are being broken, but the O2−H1 and Ir− H2 bonds are being formed in TS1/4−Ir. On the other hand, the C2O3 double bond is partially formed (1.278 Å) concomitantly with the weakening of the O3−H1 and C2−H2 bonds. In 4−Ir, the O2−H1 and Ir−H2 bonds are formed, where the bond distances are 0.992 and 1.595 Å, respectively. In this step, benzaldehyde 3 is released. The ΔG0≠ and ΔG0 values of this reaction step are 12.6 and 3.5 kcal/mol, respectively, as shown in Figure 2A. In this reaction step, the pyridonate moiety of the bpyO ligand is converted to a hydroxylpyridine moiety through the protonation of the carbonyl O2 atom, which corresponds to the aromatization of the pyridonate moiety.11 Actually, the geometry of the C5N ring of the hydroxylpyridine moiety is essentially the same as that of the free hydroxylpyridine molecule; see Figure S3 in Supporting Information for the optimized geometry of hydroxylpyridine. The Ir−N1 bond is elongated by 0.050 Å when going from 1−Ir to 4−Ir; see Figure 1. The neutral hydroxylpyridine moiety in 4−Ir is less electron-donating than the anionic pyridonate moiety in 1−Ir. Also, the Ir−H2 bond formation weakens this N1→Ir charge transfer (CT) interaction, because the hydride H2 ligand is

been reported.8 Those computational studies demonstrate that the promoterless AAD reactions take place under the cooperation of the metal center and the ligand via an aromatization/dearomatization process of the ligand. However, the influence of the aromatization/dearomatization process on the activity of the catalyst has not been discussed. Recently, Yamaguchi, Fujita, and their co-workers synthesized a more effective iridium complex Cp*Ir(bpyO)(H2O) Cat7 (Cp* = pentamethylcyclopentadienyl, bpyO = α,α′bipyridonate); see Scheme 2.7d It works well not only in the AAD reactions of aromatic alcohols but also in the AAD reactions of the aliphatic alcohols under mild conditions. One important difference of Cat7 from Cat4 to Cat6 is the presence of the bipyridonate ligand (bpyO), which is already in an active form unlike the hydroxylpyridine ligand in Cat4 to Cat6; remember that the hydroxylpyridine must undergo deprotonation before the catalytic reaction. On the other hand, the substitution of a phenanthroline-based functional ligand for the bpyO ligand , deprotonated 2,9-dihydroxy-1,10-phenanthroline (phenO), considerably decreases the activity of the catalyst Cp*Ir(phenO)(H2O) Cat8; see Scheme 2 for Cat8 and the phenO ligand. These results suggest that the activity of this kind of iridium catalyst is sensitive to the ligand and the bpyO ligand is the optimized one so far. However, the reason is not clear. Also, it is unclear whether the mechanism of the AAD reaction by Cat 7 is the same as that by previously reported catalysts such as Cat4 or not. In this work, we theoretically investigated the AAD reactions of primary and secondary alcohols with several iridium complexes to elucidate the reaction mechanism, to explore how much the aromatization/dearomatization process of the ligand influences the activity of the catalyst, and to explain why

Scheme 3. Proposed Reaction Pathways of the AAD Reaction Mediated by Cat 7

1011

dx.doi.org/10.1021/cs401101m | ACS Catal. 2014, 4, 1010−1020

ACS Catalysis

Research Article

Figure 1. Geometry and energy changes of the AAD reaction of the benzyl alcohol 2 by Cp*Ir(bpyO) 1−Ir along the outer-sphere concerted pathway. Hydrogen atoms on the Cp* and bpyO ligands are omitted for clarity. Bond distances are in angstrom. In parentheses are Gibbs energy changes (in kcal/mol).

strongly electron-donating. Due to these two factors, the Ir−N1 bond is elongated. The next step is the formation of a dihydrogen complex Cp*Ir(bpyO)(H2) 5−Ir. This process is facilitated by an alcohol bridge, which occurs through a transition state TS4/5− Ir, as shown in Figure 1. In TS4/5−Ir, the protonic H1 atom migrates from the hydroxyl O2 atom of the ligand to the hydroxyl O3 atom of the bridging alcohol molecule, where the O2−H1 and O3−H1 distances are 1.457 and 1.040 Å, respectively. Simultaneously, the protonic H3 atom migrates from the hydroxyl O3 atom to the hydride H2 ligand, where the O3−H3 and H2−H3 distances are 1.416 and 0.987 Å, respectively. In TS4/5−Ir, the Ir−H2 bond length increases to 1.676 Å, comparing with that (1.595 Å) in 4−Ir. These geometrical features indicate that the H2−H3 bond is being formed concomitantly with the Ir−H2 bond breaking in TS4/5−Ir. In 5−Ir, the Ir−H2 bond is further elongated to 1.731 Å, whereas the H2−H3 bond distance further decreases to 0.882 Å. These geometrical changes indicate that a dihydrogen complex is formed. This reaction step occurs with a ΔG0≠ value of 23.9 kcal/mol and a ΔG0 value of 13.9 kcal/ mol. The deprotonation of the pyOpyOH ligand leads to the dearomatization of the pyOH moiety, as follows: The C1−O2 bond returns back to a double bond (1.245 Å) in 5−Ir from a single bond (1.330 Å) in 4−Ir. The C1−N1 conjugated double bond simultaneously changes to a single bond (1.395 Å). Also, the C1−C2 and C3−C4 bonds become longer, but the C2−C3 and C4−C5 bonds become shorter. These features indicate that the aromatic nature of the C5N ring disappears in 5−Ir. We also investigated the transformation of 4−Ir to 5−Ir without the alcohol bridge; see Figure S4 for the geometry of the transition state.12 In this case, the ΔG0≠ value increases to

Figure 2. (A) Reaction profile for the AAD reaction of benzyl alcohol 2 mediated by Cp*Ir(bpyO) 1−Ir and Cp*Ir(pyOhpyO) 1−IrH. (B) Reaction profile for the AAD reaction of 1-phenylethanol 2′ mediated by Cp*Ir(bpyO) 1−Ir (Cat7) and Cp*Ir(phenO) 1−IrP (Cat8).

1012

dx.doi.org/10.1021/cs401101m | ACS Catal. 2014, 4, 1010−1020

ACS Catalysis

Research Article

Figure 3. Geometry and energy changes of the AAD reaction of the benzyl alcohol 2 by Cp*Ir(bpyO) 1−Ir along the inner-sphere β-H elimination pathway. Hydrogen atoms on the Cp* and bpyO ligands are omitted for clarity. Bond distances are in angstrom. In parentheses are Gibbs energy changes (in kcal/mol).

transfer process. In 7−Ir, the Ir−O3 distance is shorter by 0.106 Å than that in 6−Ir, indicating that the alkoxide moiety more strongly coordinates with the iridium center than the alcohol molecule, as expected. The ΔG0≠ and ΔG0 values of this step are 10.2 and 5.8 kcal/mol, respectively. In the third step, the β-C−H bond of the alkoxide moiety is activated through a transition state TS7/4−Ir to produce a hydride complex 4−Ir with the release of benzaldehyde 3. When going from 7−Ir to TS7/4−Ir, the Ir−O3−C2 angle significantly decreases from 125° to 78°, which makes the C2− H2 bond approach the iridium center. Actually, the Ir−H2 distance significantly decreases from 3.134 to 1.658 Å. As the C2−H2 bond approaches the iridium center, its bond distance considerably increases from 1.100 to 1.591 Å. At the same time, the C2−O3 distance somewhat decreases to 1.319 Å from 1.407 Å, indicating that the C2−O3 bond becomes stronger. This β-C−H activation step occurs with the ΔG0≠ and ΔG0 values of 42.4 and −7.8 kcal/mol, respectively. The ΔG0≠ value is too large for this reaction pathway to occur at the experimental temperature (353.15 K). Hence, this mechanism can be excluded. The above result is different from the previous computational result of the AAD reaction by Cat4, in which the inner-sphere β-H elimination pathway occurs with comparable ΔG0≠ value to that of the outer-sphere concerted pathway.8b In the active form of Cat4, a monodentate pyridonate ligand coordinates with the iridium center, which changes to a hydroxylpyridine ligand in the proton transfer step. The hydroxylpyridine ligand may easily dissociate from the iridium center to generate enough space for the following β-H elimination reaction. This is favorable for the inner-sphere β-H elimination mechanism. In Cat7, however, the bpyO lignad changes to a pyOpyOH ligand in the proton transfer step. The pyOpyOH would not completely dissociate from the iridium center, because the

31.0 kcal/mol, indicating that the direct formation of the dihydrogen complex is difficult. The last step is the H2 elimination from the iridium center through a transition state TS5/1−Ir to regenerate the active species 1−Ir. When going from 5−Ir to TS5/1−Ir, the Ir−H2 and Ir−H3 distances are significantly elongated to 2.353 and 2.353 Å, respectively, and the H2−H3 distance (0.758 Å) becomes almost the same as that (0.748 Å) of the free H2 molecule. This reaction step occurs with the ΔG0≠ and ΔG0 values of 23.0 and −4.2 kcal/mol, respectively; see Figure 2A. The outer-sphere stepwise pathway was also investigated here, where the O2−H1 and C2−H2 bonds are cleaved successively in a stepwise manner via an ion pair intermediate; see Scheme S1 in the Supporting Information. However, the ion pair intermediate could not be calculated as an equilibrium species, indicating that this mechanism is not favorable. Inner-Sphere β-H Elimination Pathway. In the first step of this pathway, the benzyl alcohol 2 coordinates with the iridium center of 1−Ir via the oxygen atom to form a precursor complex Cp*Ir(bpyO)(PhCH2OH) 6−Ir; see Scheme 3. When going from TS1/6−Ir to 6−Ir, the Ir−O3 distance becomes significantly shorter from 3.031 Å to 2.202 Å, indicating that 2 certainly coordinates with the iridium center; see Figure 3. As a result, the O3−H1 bond is activated, where the bond distance becomes longer by 0.032 Å in 6−Ir than in the free alcohol molecule 2. The ΔG0≠ and ΔG0 values of this coordination step are 9.6 and 5.5 kcal/mol, respectively. In the next step, the H1 atom moves from the hydroxyl O3 atom of 2 to the carbonyl O2 atom of the bpyO ligand, leading to an alkoxide complex Cp*Ir(pyOpyOH)(OCH2Ph) 7−Ir. In the transition state TS6/7−Ir, the O3−H1 and O2−H1 distances are 1.311 and 1.138 Å, respectively. The natural atomic charges of O2, H1, and O3 atoms are −0.738, 0.538, and −0.728 e, respectively, suggesting that this is a proton 1013

dx.doi.org/10.1021/cs401101m | ACS Catal. 2014, 4, 1010−1020

ACS Catalysis

Research Article

anionic pyridonate moiety to a neutral hydroxylpyridine moiety. Unexpectedly, the O2 atomic population changes little, even though the protonation occurs at the O2 atom. This phenomenon indicates that the protonic H1 atom is mainly provided electron population by the C5N(a) ring, which is consistent with the decrease in the electron population of the C5N(a) ring. In the formation step of the dihydrogen complex, the strongly electron-donating hydride ligand changes to a weakly electron-donating dihydrogen ligand. As a result, the iridium atomic population somewhat decreases, when going from 4−Ir to 5−Ir, as shown in Figure 4. To compensate for the decrease in the iridium atomic population, the CT starts to occur from the Cp* ligand to the iridium center. This is the origin of the decrease in the electron population of the Cp* ligand. Simultaneously, the electron population of the active C5N(a) ring somewhat increases, which corresponds to the dearomatization of the neutral hydroxylpyridine moiety to the anionic pyridonate moiety. These population changes indicate that the CT from the Cp* ligand to the iridium center contributes to the stabilization of TS4/5−Ir in the formation step of the dihydrogen complex, which is rate-determining in the AAD reaction. In the H2 elimination step, the dihydrogen molecule dissociates from the iridium center to regenerate the active species 1−Ir. In this step, the electron population of the iridium center further decreases, because the CT from the dihydrogen molecule to the iridium center gradually disappears. As a result, the CT from the Cp* ligand to the iridium center becomes stronger, which leads to a further decrease in the electron population of the Cp* ligand, as shown in Figure 4. According to the above analysis, the Cp* ligand receives electrons in the dehydrogenation process and releases them in the rate-determining step to stabilize the corresponding transition state. Hence, the electron-donating Cp* ligand is favorable for the AAD reaction. This is consistent with our calculation results that when the Cp* ligand is substituted for a weaker electron-donating Cp ligand, the Gibbs energy difference between 4−CpIr and TS4/5−CpIr (the ratedetermining step) increases to 24.3 kcal/mol; see Figure S5 for the optimized geometries and energy changes. Effect of Ligand Aromatization/Dearomatization on the Activity of Cp*Ir(bpyO) 1−Ir. To examine how much the aromatization/dearomatization process of the ligand influences the activity of 1−Ir, we employed a model complex Cp*Ir(pyOhpyO) in the AAD reaction, where the pyOhpyO ligand consists of a pyridonate ring (pyO) and a hydrogenated pyridonate (hpyO) ring, as shown in Figure 5. Cp*Ir(pyOhpyO) is named as 1−IrH, hereafter. In the following discussion, we investigated the AAD reaction which occurs under the cooperation of the iridium center and the hpyO moiety of the ligand; note that the aromatization/dearomatization process does not occur in the hpyO moiety, because the C5N ring of the hpyO moiety is saturated with hydrogen atoms. In 1−IrH, the Ir−N1 distance is somewhat shorter than that in 1−Ir by 0.028 Å (as shown in Figure 5), whereas the Ir−N2 bond distance becomes moderately longer by 0.009 Å. These results indicate that the interaction of the iridium center with the nonconjugated hpyO moiety is stronger than that with the conjugated pyO moiety. These are consistent with the fact that the electron population of the hpyO moiety decreases by 0.60 e and that of the pyO by 0.52 e when the pyOhpyO ligand coordinates with the iridium center; see Table S1 in Supporting

pyOpyOH is a chelating ligand and the remaining pyO moiety strongly coordinates with the iridium center (see Figure 3). As a result, the structure around the iridium center is congested in TS7/4−Ir. Moreover, the pyOH moiety would partially dissociate from the iridium center to keep the six-coordinate structure of the iridium center in the β-H abstraction step, which further destabilizes TS7/4−Ir; remember that the iridium center takes a d6 electron configuration. Because of these two factors, the β-H elimination pathway needs a considerably large ΔG0≠ value in the Cat7-catalyzed AAD reaction. However, the catalytic efficiency of Cat4 is lower than Cat7. This is because Cat4 must undergo the successive HCl elimination and the ligand rotation to generate the active form of the catalyst. These processes increase the energy barrier of the AAD reaction by Cat4,8b which leads to the lower catalytic efficiency of Cat4 than that of Cat7. As discussed above, the outer-sphere concerted pathway is more favorable than the inner-sphere β-H elimination one in the AAD reaction by Cat7. In the former case, the formation step of dihydrogen complex (4−Ir to 5−Ir) is the ratedetermining step with a ΔG0≠ value of 23.9 kcal/mol.13 In the latter case, the rate-determining step is the β-C−H bond activation step (7−Ir to 4−Ir) with a ΔG0≠ value of 42.4 kcal/ mol. Electron Redistribution along the Outer-sphere Concerted Pathway. In the dehydrogenation step, the H2 atom migrates from the alcohol to the iridium center to form an Ir−H2 bond, and the H1 atom moves to the carbonyl oxygen (O2) atom of the bpyO ligand to form an O2−H1 bond, as discussed above. These geometrical changes occur with considerably large population changes. The iridium atomic population considerably increases when going from 1−Ir to 4− Ir, as shown in Figure 4. This is because the CT occurs from

Figure 4. NBO population changes of Ir, Cp*, C5N(a), C5N (b), and O2 along the outer-sphere concerted pathway of the AAD reaction of benzyl alcohol 2 by Cp*Ir(bpyO) 1−Ir. C5N(a) and C5N(b) represent the active C5N and inactive C5N rings, respectively. (a) A positive value represents an increase in electron population relative to 1−Ir, and vice versa. (b) The electron population of aldehyde PhCHO does not change after here, because it releases from the reaction system in the dehydrogenation step of alcohol.

the H2 ligand to the iridium center. Such CT suppresses the CT from the Cp* ligand to the iridium center, which leads to an increase in the electron population of the Cp* ligand. On the other hand, the electron population of the active C5N(a) ring considerably decreases, whereas that of the inactive C5N(b) ring decreases little. This is because the protonation of the carbonyl O2 atom induces the aromatization of the 1014

dx.doi.org/10.1021/cs401101m | ACS Catal. 2014, 4, 1010−1020

ACS Catalysis

Research Article

Figure 5. Geometry and energy changes of all the species involved in the AAD reaction of benzyl alcohol 2 mediated by Cp*Ir(pyOhpyO) 1−IrH. Hydrogen atoms on the Cp* ligand and the pyO moiety are omitted for clarity. Bond distances are in angstrom. In parentheses are Gibbs energy changes (in kcal/mol).

Information. This is because the nitrogen lone pair orbital energy is higher in the hpyO moiety than in the pyO moiety; 1.43 eV in the hpyO and 1.22 eV in the pyO, where the Kohn− Sham orbital energies are presented. The AAD reaction mediated by 1−IrH occurs through a transition state TS1/4−IrH to afford a hydride complex 4−IrH with similar geometrical changes to those by 1−Ir, as shown in Figure 5. In this step, the H2 atom migrates from the C2 atom of benzyl alcohol 2 to the iridium center. At the same time, the H1 atom moves from the hydroxyl group of the benzyl alcohol to the carbonyl O2 atom of the hpyO moiety, leading to the formation of a pyOhpyOH ligand. In TS1/4−IrH, the C2−H2 and O3−H1 bond distances are elongated to 1.560 and 1.296 Å, respectively, whereas the Ir−H2 and O2−H1 bond distances become shorter to 1.668 and 1.114 Å, respectively. These geometrical features are similar to those in TS1/4−Ir; compare Figures 1 and 5. TS1/4−IrH is moderately more unstable than TS1/4−Ir by 0.8 kcal/mol, whereas the hydride intermediate 4− IrH lies much higher than 4−Ir by 3.3 kcal/mol, as shown in Figure 2(A). These differences indicate that the aromatization of the ligand plays an important role in stabilizing the hydride intermediate. To understand well how much the aromatization of the ligand influences the relative stabilities of 4−Ir and 4−IrH, we compared the geometries and important bond energies between these two species. In 4−IrH, both the Ir−H2 and O2−H1 bond distances and their bond energies are similar to those in 4−Ir, indicating that the ligand aromatization hardly influences these two bonds; see Scheme S2 and Table S2 in Supporting Information for bond energies. However, a

significant difference is found between the hpyOH moiety of the pyOhpyOH ligand in 4−IrH and the corresponding pyOH moiety of the pyOpyOH ligand in 4−Ir, as follows: The C1− N1 bond of the pyOhpyOH ligand becomes shorter by 0.082 Å in the alcohol dehydrogenation step, while that of the bpyO ligand becomes less shorter by 0.062 Å. This is because the πelectron is localized on the C1−N1 bond in the hpyOH moiety for the lack of conjugation; note that the π-electrons are delocalized in the conjugated pyOH ring. These results indicate that the ligand aromatization stabilizes the hydride intermediate through the delocalization of π-electrons. The next step is the formation of a dihydrogen complex 5− IrH through a transition state TS4/5−IrH. In this step, the protonic H1 atom moves from the pyOhpyOH ligand to the O3 atom of the bridging alcohol, and the protonic H3 atom of the bridging alcohol moves toward the H2 hydride ligand to form a H2−H3 bond. TS4/5−IrH resembles well TS4/5−Ir in geometry. When the most stable species (1−IrH + 2) is taken as a reference, the ΔG0≠ value (26.6 kcal/mol) of TS4/5−IrH is higher than that (23.9 kcal/mol) of TS4/5−Ir, as shown in Figure 2A. However, the energy difference (19.8 kcal/mol) between 4−IrH and TS4/5−IrH is smaller than that (20.4 kcal/ mol) between 4−Ir and TS4/5−Ir. This is because the dearomatization of the ligand occurs when going from 4−Ir to TS4/5−Ir but it does not occur when going from 4−IrH to TS4/5−IrH. These results indicate that the less stability of TS4/5−IrH than TS4/5−Ir arises from the less stability of 4− IrH than 4−Ir (by 3.3 kcal/mol). In brief, the aromatization of the ligand directly stabilizes the hydride intermediate 4−Ir, which in turn leads to the stabilization of the transition state in 1015

dx.doi.org/10.1021/cs401101m | ACS Catal. 2014, 4, 1010−1020

ACS Catalysis

Research Article

(bpyO) 1−Rh in the AAD reaction of 1-phenylethanol 2′; here the secondary alcohol was employed as a substrate, because we were afraid that 1−Rh was less active than 1−Ir. The AAD reaction by 1−Rh occurs with similar geometry and energy changes to those of the reaction by 1−Ir; see Figure S7 in Supporting Information. The rate-determining step is the formation of the dihydrogen complex like that in the catalytic cycle by 1−Ir, where the ΔG0≠ and ΔG0 values are 21.2 and 12.4 kcal/mol, respectively, relative to the sum of 1−Rh and 2′. These ΔG0≠ and ΔG0 values are similar to those of the AAD reaction of 2′ mediated by 1−Ir (22.3 and 13.8 kcal/mol, respectively); see Figure S8. These calculation results indicate that 1−Rh would be a good candidate for the AAD reaction, which is consistent with the experimental results reported below. Encouraged by the good result about the rhodium catalyst, we tried to use other 4d metal elements such as ruthenium in the catalyst. The ruthenium element is less expensive and more abundant than the iridium element, which is widely employed in the hydrogen transfer catalyst.7f,h,14 However, the simple substitution of the iridium(III) center for a ruthenium(II) leads to the formation of a negatively charged ruthenium(II) complex, as shown in Scheme 4. Because the solubility to

the subsequent formation of a dihydrogen complex (ratedetermining step). Thus, the aromatization of the ligand is favorable for the AAD reaction. The last step is the H2 elimination from the iridium center to afford the active species 1−IrH. This process occurs in a similar way in both the bpyO and pyOhpyO systems. We wish to skip the discussion about this step. Comparison between Activities of Cp*Ir(bpyO) 1−Ir and Cp*Ir(phenO) 1−IrP. The Fujita−Yamaguchi group reported that the catalytic activity of the iridium complex becomes much lower by the substitution of a phenO ligand for the bpyO ligand .7d To elucidate the reason, we investigated the AAD reaction of 1-phenylethanol 2′ mediated by Cp*Ir(phenO) 1−IrP, which corresponds to Cat8 reported in the experimental work.7d Our calculations indicate that this reaction occurs through the same reaction mechanism as that of the AAD reaction mediated by 1−Ir via similar geometrical changes. Hence, we skipped the discussion of geometrical changes along the reaction pathway for brevity; see Figure S6 in Supporting Information for optimized geometries. As shown in Figure 2B, the rate-determining step of the AAD reaction by 1−IrP is the formation of the dihydrogen complex, as well. Its transition state TS4/5−IrP lies above TS′4/5−Ir by 3.4 kcal/mol. This is consistent with the experimental fact that the use of the phenO ligand decreases the activity of the iridium catalyst. It should be noted that the Gibbs energy difference between 4−IrP and TS4/5−IrP is smaller than that between 4− Ir and TS′4/5−Ir, where 4−IrP lies above 4−Ir by 4.5 kcal/ mol; see Figure 2B. These results indicate that the dehydrogenation process becomes difficult when the Cp*Ir(phenO) 1−IrP is employed as a catalyst, because the intermediate 4−IrP is less stable than 4−Ir. This is similar to the reaction mediated by Cp*Ir(pyOhpyO) 1−IrH; compare the energy changes in Figures 2A,B. It is necessary to explore the reasons why 4−IrP is less stable than 4−Ir. One reason is the aromatization effect, which is smaller in the phenO ligand than in the bpyO ligand, because the phenO ligand is more conjugated than bpyO due to the presence of a conjugated six-membered ring between two pyridonate (C5N) rings. Another reason is that the proton affinity (260.3 kcal/mol) of the phenO ligand is smaller than that (263.7 kcal/mol) of the bpyO ligand by about 3.4 kcal/ mol, which is similar to the energy difference between 4−Ir and 4−IrP. The smaller proton affinity of the phenO ligand in 1− IrP comes from the lower orbital energy of the oxygen lone pair (−6.91 eV) than that of the bpyO ligand (−6.77 eV) in 1−Ir; see Scheme S3 in Supporting Information. It is likely that the phenO ligand stabilizes the oxygen lone-pair orbital more than the bpyO ligand for its larger conjugate electronic structure. As a result, the phenO ligand is less reactive for the protonation reaction than the bpyO ligand, which leads to the less stable intermediate 4−IrP. In both reasons, the presence of one more six-membered ring in the phenO ligand is the key factor for the lower activity of 1−IrP. Prediction of New Catalysts. The prediction of new catalysts was attempted here on the basis of the combination of experimental and theoretical works. The theoretical result will be discussed here first, and the experimental results will be presented below. The simple and direct way to design a new catalyst with the same ligands as those in Cp*Ir(bpyO) 1−Ir is to replace the iridium center with its 4d analogue, which is the rhodium element. We examined the reactivity of the complex Cp*Rh-

Scheme 4. Prediction of a New Catalyst

organic solvent is important, a neutral ruthenium(II) complex would be better. This requires to substitute some neutral threecoordinate ligand for the monoanionic Cp* ligand; remember that the ruthenium(II) center is in d6 electron configuration, which tends to take a six-coordinate octahedral-like structure. One of the candidates is benzene, which is employed in the ruthenium catalyst for transfer hydrogenation reaction.15 According to the above discussion, the electron-donating ligand is more favorable for the AAD reaction. Therefore, we use hexamethylbenzene (HMB) instead of benzene as a ligand to construct a ruthenium complex; see Scheme 4. The activity of the new ruthenium(II) complex (HMB)Ru(bpyO) 1−Ru is examined by the AAD reaction of the 1-phenylethanol 2′. In the first step, the AAD reaction occurs with the cooperation of the ruthenium center and the bpyO ligand to afford a hydride intermediate (HMB)RuH(pyOpyOH) 4−Ru through a transition state TS1/4−Ru; see Figure 6. In TS1/4− Ru, the H2 atom is moving from the C2 atom to the ruthenium center. Simultaneously, the protonic H1 atom is moving from the O3 atom of the 1-phenylethanol to the carbonyl O2 atom of the bpyO ligand. The geometrical changes of this dehydrogenation step are similar to those of the reaction by 1−Ir. However, this step much more easily occurs by 1−Ru with a ΔG0≠ value of 1.6 kcal/mol than that by 1−Ir (ΔG0≠ = 12.6 kcal/mol). In the following, 4−Ru is converted to a dihydrogen complex (HMB)Ru(bpyO)(H2) 5−Ru through an alcohol-bridged transition state TS4/5−Ru, which is similar to TS′4/5−Ir. This step occurs with the ΔG0≠ and ΔG0 values of 16.7 and 7.9 kcal/mol, respectively, where the sum of 4−Ru and 2′ is taken as a reference (energy zero); see Figure 6. This 1016

dx.doi.org/10.1021/cs401101m | ACS Catal. 2014, 4, 1010−1020

ACS Catalysis

Research Article

Figure 6. Geometry and energy changes of all the species involved in the AAD reaction of 1-phenylethanol 2′ mediated by (HMB)Ru(bpyO) 1−Ru. Hydrogen atoms on the HMB and bpyO ligands are omitted for clarity. Bond distances are in angstrom. In parentheses are Gibbs energy changes (in kcal/mol).

ΔG0≠ value is moderately smaller than that (22.3 kcal/mol) for TS′4/5−Ir. The last step is the H2 elimination from the ruthenium center in 5−Ru, which occurs through a transition state TS5/1−Ru to regenerate the active species 1−Ru. This step needs a moderately smaller ΔG0≠ value (15.2 kcal/mol) than that of the rate-determining step like those in the reactions by 1−Ir and 1−Rh. The population changes are also similar to those of the iridium reaction system; see Figure S9 in Supporting Information. These results suggest that the AAD reaction by 1−Ru occurs with similar electronic processes to those by 1−Ir. On the basis of the geometrical and energy changes along the AAD reaction, we reached the conclusion that 1−Ru would be a good catalyst for the AAD reaction. The experimental results of 1−Ru will be presented below. Experiments of AAD Reaction with Cp*Rh(bpyO)(H2O) Cat9 and (HMB)Ru(bpyO)(H2O) Cat10. In order to carry out the experimental study on the theoretically predicted catalyst, we first attempted the synthesis of new catalysts Cp*Rh(bpyO)(H2O) Cat9 and (HMB)Ru(bpyO)(H2O) Cat10. The rhodium Cat9 was synthesized in good yield starting with [Cp*Rh(H2O)3](OTf)2 by a procedure similar to that for the iridium Cat7; see the experimental section. The ruthenium Cat10 could also be synthesized in a reasonable yield starting with [(HMB)Ru(H2O)3](OTf)2 via [(HMB)Ru(6,6′-dihydroxy-2,2′-bipyridine)(H2O)](OTf)2. Then, we performed experiments of the AAD reaction catalyzed by Cat9 and Cat10, using 1-phenylethanol as a substrate. For comparison, a similar reaction catalyzed by the previous iridium catalyst, Cp*Ir(bpyO)(H2O) Cat7, was also carried out. Results are summarized in Table 1. When the solution of 1-phenylethanol in benzene (boiling point: 80 °C) was refluxed for 20 h in the presence of 0.50 mol % of Cat7, acetophenone was obtained quantitatively along with the evolution of hydrogen gas (entry

Table 1. Catalytic Performances of Cat7, Cat9, and Cat10 in the AAD Reaction of 1-Phenylethanol under Reflux in Benzenea

entry

catalyst

conversion (%)b

yield (%)b

1 2 3

Cat7 Cat9 Cat10

100 95 97

100 95 97

a

The reaction was carried out with catalyst (0.50 mol %) and 1phenylethanol (1.0 mmol) in benzene (3 mL) under reflux for 20 h. b Determined by GC.

1).16,17a Analogous rhodium Cat9 also showed high catalytic activity to give acetophenone in 95% yield (entry 2), supporting the theoretical prediction described above. The AAD reaction catalyzed by ruthenium Cat10 resulted in 97% yield of acetophenone (entry 3), also in accord with the theoretical prediction.17b

■

CONCLUSIONS In this work, we theoretically investigated the promoterless AAD reactions mediated by Cp*Ir(bpyO) 1−Ir developed by Yamaguchi, Fujita, and their co-workers, and we theoretically 1017

dx.doi.org/10.1021/cs401101m | ACS Catal. 2014, 4, 1010−1020

ACS Catalysis

Research Article

Hence, the B3PW91 functional was employed in all the calculations in this work.

and experimentally found new catalysts with the rhodium and ruthenium elements. The AAD reaction of benzyl alcohol 2 by 1−Ir occurs through three steps: alcohol dehydrogenation, formation of dihydrogen complex, and elimination of dihydrogen molecule. In the first step, the outer-sphere concerted pathway is more favorable than the inner-sphere β-H elimination pathway, which is different from the reaction mechanism of the AAD reaction by a Cp*Ir(pyridonate) complex Cat4 (see Scheme 2). The second step is facilitated by an alcohol bridge. In these two steps, the metal center and the bpyO ligand work cooperatively via the aromatization and dearomatization processes of the ligand. The second step is rate-determining, where the ΔG0≠ and ΔG0 values are 23.9 and 13.9 kcal/mol, respectively. When the pyOhpyO ligand is employed in the catalyst instead of the bpyO ligand, the ΔG0≠ value of the ratedetermining step becomes higher to 26.6 kcal/mol. This is because one hpyO ring of the pyOhpyO ligand is saturated with hydrogen atoms and the aromatization/dearomatization of the ligand does not occur in the AAD reaction. When the bpyO ligand is substituted for a larger conjugated phenO ligand, the rate-determining step also becomes more difficult; the ΔG0≠ value is 25.7 kcal/mol. This result arises from the smaller proton affinity of the oxygen atom and smaller aromatization effect of the phenO ligand. We proposed new AAD catalysts, Cp*Rh(bpyO) 1−Rh and (HMB)Ru(bpyO) 1−Ru, with the combination of the theoretical results and experimental intuition. Both theoretical calculations and experimental results show that these complexes are good catalysts for the AAD reaction. The AAD reaction by 1−Rh occurs with similar geometry and energy changes to those by 1−Ir, where the ΔG0≠ value of the rate-determining step is 21.2 kcal/mol. The AAD reaction by 1−Ru occurs with similar ΔG0≠ value of 17.0 kcal/mol for the rate-determining step. These results strongly suggest that the main structure of M(bpyO) is crucial and can be applied to a new catalyst for the AAD reactions.

■

■

EXPERIMENTAL DETAILS

1

H and 13C NMR spectra were recorded on JEOL ECX-500 and ECS400 spectrometers. Gas chromatography (GC) analyses were performed on a GL-Sciences GC353B gas chromatograph with a capillary column (GL-Sciences InertCap Pure WAX). Elemental analyses were carried out at the Microanalysis Center of Kyoto University. [Cp*Rh(H2O)3](OTf)2,30 6,6′-dihydroxy-2,2′-bipyridine,31 and Cat77d were prepared according to the literature method. Organic solvents were dried by standard procedures prior to use. All other reagents are commercially available and were used as received. Preparation of [Cp*Rh(6,6′-dihydroxy-2,2′-bipyridine)(H2O)](OTf)2. Under an atmosphere of argon, [Cp*Rh(H2O)3](OTf)2 (0.384 g, 0.65 mmol) was placed in a flask. Water (13 mL) and 6,6′-dihydroxy-2,2′-bipyridine (0.185 g, 0.98 mmol) were added, and the mixture was stirred for 1 h at room temperature. After the solution was filtered through a pad of Celite, the solvent was removed in a vacuum to gave a yellow powder of the title product in 84% yield (0.408 g, 0.55 mmol). 1H NMR (400 MHz, D2O): δ 8.04 (t, J = 8 Hz, 2H, py), 7.83 (d, J = 7 Hz, 2H, py), 7.17 (d, J = 8 Hz, 2H, py), 1.58 (s, 15H, Cp*). 13C NMR (100 MHz, D2O): δ 164.5 (s, py), 154.1 (s, py), 144.0 (s, py), 120.3 (q, JC−F =317 Hz, OTf), 116.1 (s, py), 114.0 (s, py), 98.1 (d, JC−Rh = 10 Hz, C5Me5), 9.2 (s, C5Me5). Preparation of Cp*Rh(bpyO)(H2O) Cat9. Under an atmosphere of argon, [Cp*Rh(6,6′-dihydroxy-2,2′-bipyridine)(H2O)](OTf)2 (0.445 g, 0.60 mmol) was placed in a flask. Water (25 mL) and NaOtBu (0.112 g, 1.2 mmol) were added, and the mixture was stirred for 30 min at room temperature. The yellow solution gradually changed to orange suspension. The suspension was filtered and washed with water (6 mL) and diethylether (5 mL). Drying under vacuum gave Cat9 as an orange powder (0.238 g, 0.54 mmol, 90%). 1 H NMR (400 MHz, CD3OD): δ 7.44 (t, J = 8 Hz, 2H, py), 6.96 (d, J = 7 Hz, 2H, py), 6.44 (d, J = 9 Hz, 2H, py), 1.58 (s, 15H, Cp*). 13C NMR (100 MHz, CD3OD): δ 171.7 (s, py), 157.2 (s, py), 140.1 (s, py), 117.9 (s, py), 107.4 (s, py), 96.8 (d, JC−Rh = 9 Hz, C5Me5), 9.6 (s, C5Me5). Anal. Calcd for C20H23N2O3Rh: C, 54.31; H, 5.24; N, 6.33. Found: C, 53.77; H, 5.01; N, 6.33. Preparation of [(HMB)Ru(H2O)3](OTf)2. Under an atmosphere of argon, [(HMB)RuCl2]2 (0.993 g, 1.5 mmol) was placed in a flask. Water (30 mL) and AgOTf (1.54 g, 6.0 mmol) were added, and the mixture was stirred for 4 h at room temperature. After filtration through a pad of Celite, evaporation of the filtrate gave the title product as an orange powder (1.70 g, 2.9 mmol, 99%). 1H NMR (400 MHz, D2O): δ 2.19 (s, C6Me6). 13C NMR (100 MHz, D2O): δ 120.3 (q, JC−F = 317 Hz), 89.7 (s, C6Me6), 15.8 (s, C6Me6). Preparation of [(HMB)Ru(6,6′-dihydroxy-2,2′-bipyridine)(H2O)](OTf)2. Under an atmosphere of argon, [(HMB)Ru(H2O)3](OTf)2 (1.78 g, 3.0 mmol) was placed in a flask. Water (30 mL) and 6,6′-dihydroxy-2,2′-bipyridine (0.594 g, 3.2 mmol) were added, and the mixture was stirred for 2.5 h at room temperature. After the solution was filtered through a pad of Celite, the solvent was removed in a vacuum to give an orange powder of the title product in 77% yield (1.77 g, 2.3 mmol). 1H NMR (500 MHz, D2O): δ 8.00 (t, J = 8 Hz, 2H, py), 7.81 (d, J = 8 Hz, 2H, py), 7.13 (d, J = 8 Hz, 2H, py), 1.99 (s, 18H, C6Me6). 13C NMR (125 MHz, D2O): δ 166.2 (s, py), 154.6 (s, py), 143.5 (s, py), 120.2 (q, JC−F = 317 Hz), 116.1 (s, py), 113.5 (s, py), 94.7 (s, C6Me6), 16.1 (s, C6Me6). Preparation of (HMB)Ru(bpyO)(H2O) (Cat10). Under an atmosphere of argon, [(HMB)Ru(6,6′-dihydroxy-2,2′-bipyridine)(H2O)](OTf)2 (0.389 g, 0.51 mmol) was placed in a flask. Water (10 mL) and NaOtBu (0.0983 g, 1.0 mmol) were added, and the mixture was stirred for 1 h at room temperature. The orange solution gradually changed to a green suspension. The suspension was filtered and washed with diethylether (10 mL). Drying under vacuum gave Cat10 as a green powder (0.192 g, 0.41 mmol, 82%). 1H NMR (400 MHz, CD3OD): δ 7.41 (t, J = 8 Hz, 2H, py), 6.95 (d, J = 7 Hz, 2H, py), 6.39 (d, J = 8 Hz, 2H, py), 1.99 (s, 18H, C6Me6). 13C NMR (100 MHz, CD3OD): δ 173.6 (s, py), 157.0 (s, py), 139.7 (s, py), 117.2 (s,

COMPUTATIONAL DETAILS

Geometry optimizations were performed by the density functional theory (DFT)18 with the B3PW91 functional19 in the gas phase. The effective core potentials (ECPs) of the Stuttgart−Dresden−Bonn group20 were employed for the core electrons of the iridium, rhodium, and ruthenium elements. The (311111/22111/411/11) basis sets were used for their valence electrons.21 The 6-31+G* basis sets22 were employed for the Cp* group and the phenyl group of benzyl alcohol and 1-phenylethanol. For other atoms, 6-31++G** basis sets22 were used. Frequency calculation was carried out for each stationary structure to make sure whether it is an equilibrium structure or a transition state. The solvent effect of benzene was evaluated with the conductor-like polarizable continuum model (CPCM),23 where optimized structures in gas phase were employed. The Gibbs energy in solution was employed for discussion, where the translational entropy was evaluated with the method developed by Whitesides et al.,24 similar to our previous works.25 The natural bond orbital analysis was performed for the electron population. All these calculations were carried out by the Gaussian 09 program.26 The reliability of the method employed in this work was evaluated by the calculations of the AAD reaction of ethanol by Cp*Ir(bpyO) 1−Ir. The geometries were optimized by the B3PW91 functional in the gas phase; see Figure S1 in Supporting Information. The energy barrier of the rate-determining step was evaluated by the SCS-MP2,27 MP228 to MP4SDQ29 methods in solution phase with the CPCM model. As shown in Table S3 in Supporting Information, the B3PW91-calculated results are close to the MP4SDQ-calculated ones. 1018

dx.doi.org/10.1021/cs401101m | ACS Catal. 2014, 4, 1010−1020

ACS Catalysis

Research Article

H. G. L.; Gusevskaya, E. V. J. Mol. Catal. A: Chem. 2000, 164, 39−47. (h) Royer, A. M.; Rauchfuss, T. B.; Wilson, S. R. Inorg. Chem. 2008, 47, 395−397. (i) Baratta, W.; Bossi, G.; Putignano, E.; Rigo, P. Chem.Eur. J. 2011, 17, 3474−3481. (j) Prades, A.; Peris, E.; Albrecht, M. Organometallics 2011, 30, 1162−1167. (k) Musa, S.; Shaposhnikov, I.; Cohen, S.; Gelman, D. Angew. Chem., Int. Ed. 2011, 50, 3533−3537. (4) Sieffert, N.; Bühl, M. J. Am. Chem. Soc. 2010, 132, 8056−8070. (5) (a) Zhang, J.; Balaraman, E.; Leitus, G.; Milstein, D. Organometallics 2011, 30, 5716−5724. (b) Zhang, J.; Leitus, G.; Ben-David, Y.; Milstein, D. J. Am. Chem. Soc. 2005, 127, 10840− 10841. (c) Gunanathan, W.; Shimon, L. J. W.; Milstein, D. J. Am. Chem. Soc. 2009, 131, 3146−3147. (d) Balaraman, E.; Gnanaprakasam, B.; Shimon, L. J. W.; Milstein, D. J. Am. Chem. Soc. 2010, 132, 16756− 16758. (e) Zhang, J.; Gandelman, M.; Shimon, L. J. W.; Rozenberg, H.; Milstein, D. Organometallics 2004, 23, 4026−4033. (f) Srimani, D.; Balaraman, E.; Gnanaprakasam, B.; Ben-David, Y.; Milstein, D. Adv. Synth. Catal. 2012, 354, 2403−2406. (g) Gunanathan, C.; Milstein, D. Acc. Chem. Res. 2011, 44, 588−602. (6) (a) Zhang, G.; Vasudevan, K. V.; Scott, B. L.; Hanson, S. K. J. Am. Chem. Soc. 2013, 135, 8668−8681. (b) Zhang, G.; Hanson, S. K. Org. Lett. 2013, 15, 650−653. (7) (a) Fujita, K.; Tanino, N.; Yamaguchi, R. Org. Lett. 2007, 9, 109− 111. (b) Yamaguchi, R.; Ikeda, C.; Takahashi, Y.; Fujita, K. J. Am. Chem. Soc. 2009, 131, 8410−8412. (c) Fujita, K.; Yoshida, T.; Imori, Y.; Yamaguchi, R. Org. Lett. 2011, 13, 2278−2281. (d) Kawahara, R.; Fujita, K.; Yamaguchi, R. J. Am. Chem. Soc. 2012, 134, 3643−3646. (e) Kawahara, R.; Fujita, K.; Yamaguchi, R. Angew. Chem., Int. Ed. 2012, 51, 12790−12794. (f) Nieto, I.; Livings, M. S.; Sacci, J. B. I.; Reuther, L. E.; Zeller, M.; Papish, E. T. Organometallics 2011, 30, 6339−6342. (g) Hull, J. F.; Himeda, Y.; Wang, W.-H.; Hashiguchi, B.; Periana, R.; Szalda, D. J.; Muckerman, J. T.; Fujita, E. Nat. Chem. 2012, 4, 383−388. (h) Moore, C. M.; Szymczak, N. K. Chem. Commun. 2012, 49, 400−402. (i) DePasquale, J.; Nieto, I.; Reuther, L. E.; Herbst-Gervasoni, C. J.; Paul, J. J.; Mochalin, V.; Zeller, M.; Thomas, C. M.; Addison, A. W.; Papish, E. T. Inorg. Chem. 2013, 52, 9175− 9183. (j) Badiei, Y. M.; Wang, W.-H.; Hull, J. F.; Szalda, D. J.; Muckerman, J. T.; Himeda, Y.; Fujita, E. Inorg. Chem. 2013, 52, 12576−12586. (8) (a) Zeng, G.; Li, S. Inorg. Chem. 2011, 50, 10572−10580. (b) Li, H.; Lu, G.; Jiang, J.; Huang, F.; Wang, Z. Organometallics 2011, 30, 2349−2363. (c) Li, H.; Wang, X.; Huang, F.; Lu, G.; Jiang, J.; Wang, Z. Organometallics 2011, 30, 5233−5247. (d) Li, H.; Wang, X.; Wen, M.; Wang, Z. Eur. J. Inorg. Chem. 2012, 5011−5020. (e) Li, H.; Wang, Z. Sci. China: Chem. 2012, 55, 1991−2008. (f) Li, H.; Jiang, J.; Lu, G.; Huang, F.; Wang, Z. Organometallics 2011, 30, 3131−3141. (9) The H2O ligand of the iridium catalyst Cp*Ir(bpyO)(H2O) Cat7 is believed to dissociate from the iridium center prior to the AAD reaction. The discussion of the H2O dissociation step is skipped for brevity in the text, because this process easily occurs and does not directly relate to the AAD reaction; see Figure S2 in SI for energy and geometry changes. (10) We have attempted the stoichiometric reactions of Cat7 with benzyl alcohol and 1-phenylethanol in order to isolate 4−Ir, 5−Ir, or 7−Ir. However, isolation of such species was not successful so far. (11) Lu, G.; Li, H.; Zhao, L.; Huang, F.; Schleyer, P. v. R.; Wang, Z.X. Chem.Eur. J. 2011, 17, 2038−2043. (12) We have also considered the possibility that water originated from the aquo ligand in Cat7 serves as a bridging molecule to facilitate the transformation of 4−Ir into 5−Ir. The present calculations show that the ΔG0≠ value is almost the same between the water bridge and the alcohol bridge. If water concentration is enough, both water and alcohol participate in the reaction. This possibility was also examined by the experiment; see also ref 16. (13) The value of kinetic isotope effect (KIE) in the AAD reaction of 1-phenylethanol catalyzed by Cat7 was experimentally determined to be 1.27. Such a small KIE value suggests that the rate-determining step would not include the cleavage of the C−H bond, supporting the theoretically predicted mechanism; see also ref 17.

py), 107.6 (s, py), 94.4 (s, C6Me6), 16.4 (s, C6Me6). Anal. Calcd for C22H26N2O3Ru·(H2O)3: C, 50.66; H, 6.18; N, 5.37. Found: C, 50.49; H, 6.06; N, 5.36. General Procedure for the AAD Reactions Using Cat7, Cat9, and Cat10 Shown in Table 1. Under argon atmosphere, the catalyst (0.50 mol %), benzene (3 mL), and 1-phenylethanol (1.0 mmol) were placed in a flask. The mixture was stirred under vigorous reflux for 20 h. The conversion of 1-phenylethanol and the yield of acetophenone were determined by GC analysis using biphenyl as an internal standard.

■

ASSOCIATED CONTENT

S Supporting Information *

Optimized geometries of all the species involved in the AAD reaction of 1-phenylethanol 2′ mediated by Cp*Ir(phenO) 1− IrP and Cp*Rh(bpyO) 1−Rh, the transition state for the dihydrogen complex formation step in the AAD reaction of ethanol by 1-Ir, the species in the H2O dissociation step from Cat7, free hydroxypyridine, and the transition state for the direct conversion from 4−Ir to 5−Ir. Population changes of Ir, Cp*, C5N(a), C5N(b), and O2 along the concerted pathway of the AAD reaction of 2′ by Cp*Ir(bpyO) 1−Ir and 1−Ru. The lone pair orbital energies on the O atom of Cp*Ir(bpyO) 1−Ir, Cp*Ir(phenO) 1−IrP, and (HMB)Ru(bpyO) 1−Ru. The Cartesian coordinates of all the species in this work. This material is available free of charge via the Internet at http:// pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. *E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS S.S. acknowledges financial support by the Grants-in-Aid from the Ministry of Education, Culture, Science, Sport, and Technology through Grants-in-Aid of Specially Promoted Science and Technology (no. 22000009). S.S. and G.Z. are also thankful to the computational facility at the Institute of Molecular Science, Okazaki, Japan. K.F. thanks financial support by the Grant-in-Aid for Scientific Research on Innovative Areas (no. 25105732).

■

REFERENCES

(1) (a) Dobereiner, G. E.; Crabtree, R. H. Chem. Rev. 2010, 110, 681−70. (b) Cheng, C.; Hong, S. H. Org. Biomol. Chem. 2011, 9, 20− 26. (c) Gunanathan, C.; Milstein, D. Science 2013, 341, 1229712−1− 1229712−1. (2) (a) Friedrich, A.; Schneider, S. ChemCatChem 2009, 1, 72−73. (b) Johnson, T. C.; Morris, D. J.; Wills, M. Chem. Soc. Rev. 2010, 39, 81−88. (c) Junge, H.; Loges, B.; Beller, M. Chem. Commun. 2007, 522−524. (d) Nielsen, M.; Kammer, A.; Cozzula, D.; Junge, H.; Gladiali, S.; Beller, M. Angew. Chem., Int. Ed. 2011, 50, 9593−9597. (3) (a) Dobson, A.; Robinson, S. D. J. Organomet. Chem. 1975, 87, C52−C53. (b) Dobson, A.; Robinson, S. D. Inorg. Chem. 1977, 16, 137−142. (c) Morton, D.; Cole-Hamilton, D. J. J. Chem. Soc., Chem. Commun. 1987, 248−249. (d) Morton, D.; Cole-Hamilton, D. J. J. Chem. Soc., Chem. Commun. 1988, 1154−1156. (e) Ligthart, G. B. W. L.; Meijer, R. H.; Donners, M. P. J.; Meuldijk, J.; Vekemans, J. A. J. M.; Hulshof, L. A. Tetrahedron Lett. 2003, 44, 1507−1509. (f) Adair, G. R. A.; Williams, J. M. J. Tetrahedron Lett. 2005, 46, 8233−8235. (g) Robles-Dutenhefner, P. A.; Moura, E. M.; Gama, G. J.; Siebald, 1019

dx.doi.org/10.1021/cs401101m | ACS Catal. 2014, 4, 1010−1020

ACS Catalysis

Research Article

(14) (a) Hashiguchi, S.; Noyori, R. Acc. Chem. Res. 1997, 30, 97−102. and references therein. (b) Murata, K.; Ikariya, T.; Noyori, R. J. Org. Chem. 1999, 64, 2186−2187. (c) Mashima, K.; Abe, T.; Tani, K. Chem. Lett. 1998, 1201−1202. (d) Mashima, K.; Abe, T.; Tani, K. Chem. Lett. 1998, 1199−1200. (e) Jiang, Y.; Jiang, Q.; Zhu, G.; Zhang, X. Tetrahedron Lett. 1997, 38, 215−218. (f) Palmer, M.; Walsgrove, T.; Wills, M. J. Org. Chem. 1997, 62, 5226−5228. (g) Alonso, D. A.; Guijarro, D.; Pinho, P.; Temme, O.; Andersson, P. G. J. Org. Chem. 1998, 63, 2749−2751. (h) Jiang, Y.; Jiang, Q.; Zhang, X. J. Am. Chem. Soc. 1998, 120, 3817−3818. (15) (a) Carmona, D.; Viguri, F.; Lamata, M. P.; Ferrer, J.; Bardají, E.; Lahoz, F. J.; García-Orduña, P.; Oro, L. A. Dalton Trans. 2012, 41, 10298−10308. (b) Ogo, S.; Abura, T.; Watanabe, Y. Organometallics 2002, 21, 2964−2969. (c) Ogo, S.; Kabe, R.; Hayashi, H.; Harada, R.; Fukuzumi, S. Dalton Trans. 2006, 4657−4663. (d) Bennett, M. A.; Mitchell, T. R. B.; Stevens, M. R.; Willis, A. C. Can. J. Chem. 2001, 79, 655−669. (e) Clapham, S. E.; Hadzovic, A.; Morris, R. H. Coord. Chem. Rev. 2004, 248, 2201−2237. (16) When the AAD reaction of 1-phenylethanol catalyzed by Cat7 (0.5 mol %) was performed in the presence of water (5.0 mol%) under reflux in benzene for 10 min, acetophenone was formed in 70% yield. This result was exactly the same as that of the reaction in the absence of water (yield of acetophenone was 70%). This experimental result is consistent with the computational results (ref 12). (17) (a) We have also carried out the AAD reaction catalyzed by Cat7 (0.05 mol%) under reflux in benzene for 20 min using a 1:1 mixture of PhCH(OH)Me and PhCD(OH)Me as starting materials. Conversions of PhCH(OH)Me and PhCD(OH)Me were 38 and 30%, respectively, whereas the yield of acetophenone was 34%. By this result, the value of the kinetic isotope effect was determined to be 1.27. (b) Under the same reaction conditions, (p-cymene)Ru(bpyO)(H2O) exhibits a much lower activity (7% yield) than that (97% yield) by (HMB)Ru(bpyO)(H2O) Cat10. (18) (a) Hohenberg, P.; Kohn, W. Phys. Rev. 1964, 136, B864−B871. (b) Kohn, W.; Sham, L. J. Phys. Rev. 1965, 140, A1133−A1138. (19) (a) Becke, A. D. Phys. Rev. A 1988, 38, 3098−3100. (b) Becke, A. D. J. Chem. Phys. 1993, 98, 5648−5652. (c) Perdew, J. P. In Electronic Structure of Solids, 91th ed.; Ziesche, P., Eschrig, H., Eds.; Akademie Verlag: Berlin, 1991. (d) Burke, K.; Perdew, J. P.; Wang, Y. In Electronic Density Functional Theory: Recent Progress and New Directions; Dobson, J. F.; Vignale, G.; Das, M. P., Ed.; Plenum: New York, 1998. (20) (a) Andrae, D.; Haeussermann, U.; Dolg, M.; Stoll, H.; Preuss, H. J. Theor. Chim. Acta 1990, 77, 123−141. (b) Bergner, A.; Dolg, M.; Kuechle, W.; Stoll, H.; Preuss, H. Mol. Phys. 1993, 80, 1431−1441. (21) Dolg, M.; Wedig, U.; Stoll, H.; Preuss, H. J. Chem. Phys. 1987, 86, 866−872 and see http://www.tc.uni-koeln.de/cgi-bin/pp. pl?language=en,job=getrefs for a complete reference list. (22) (a) Rassolov, V. A.; Pople, J. A.; Ratner, M. A.; Windus, T. L. J. Chem. Phys. 1998, 109, 1223−1229. (b) Clark, T.; Chandrasekhar, J.; Spitznagel, G. W.; Schleyer, P. v. R. J. Comput. Chem. 1983, 4, 294− 301. (c) Frisch, M. J.; Pople, J. A.; Binkley, J. S. J. Chem. Phys. 1984, 80, 3265−3269. (23) (a) Tomasi, J.; Mennucci, B.; Cammi, R. Chem. Rev. 2005, 105, 2999−3093. (b) Barone, V.; Cossi, M. J. Phys. Chem. A 1998, 102, 1995−2001. (c) Cossi, M.; Rega, N.; Scalmani, G.; Barone, V. J. Comput. Chem. 2003, 24, 669−681. (24) Mammen, M.; Shakhnovich, E. I.; Deutch, J. M.; Whitesides, G. M. J. Org. Chem. 1998, 63, 3821−3830. (25) (a) Ishikawa, A.; Nakao, Y.; Sato, H.; Sakaki, S. Inorg. Chem. 2009, 48, 8154−8163. (b) Zeng, G.; Sakaki, S. Inorg. Chem. 2011, 50, 5290−5297. (c) Zeng, G.; Sakaki, S. Inorg. Chem. 2012, 51, 4597− 4605. (d) Aono, S.; Sakaki, S. J. Phys. Chem. B 2012, 116, 13045− 13062. (26) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima,

T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, N. J.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, Ö .; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09, revision B.01; Gaussian, Inc.: Wallingford, CT, 2009. (27) Grimme, S. J. Chem. Phys. 2003, 118, 9095−9102. (28) (a) Trucks, G. W.; Watts, J. D.; Salter, E. A.; Bartlett, R. J. Chem. Phys. Lett. 1988, 153, 490−495. (b) Trucks, G. W.; Salter, E. A.; Sosa, C.; Bartlett, R. J. Chem. Phys. Lett. 1988, 147, 359−366. (29) Civalleri, B.; Zicovich-Wilson, C. M.; Valenzano, L.; Ugliengo, P. CrystEngComm 2008, 10, 405−410. (30) Eisen, M. S.; Haskel, A.; Chen, H.; Olmstead, M. M.; Smith, D. P.; Maestre, M. F.; Fish, R. H. Organometallics 1995, 14, 2806−2812. (31) Umemoto, T.; Nagayoshi, M.; Adachi, K.; Tomizawa, G. J. Org. Chem. 1998, 63, 3379−3385.

1020

dx.doi.org/10.1021/cs401101m | ACS Catal. 2014, 4, 1010−1020